[XLA] Remove note about what implementations do for DynamicSlice and DynamicUpdateSlice.

It is impossible to commit to a particular "implementation-defined behavior" for all implementations.

PiperOrigin-RevId: 190981804

diff --git a/tensorflow/docs_src/performance/xla/operation_semantics.md b/tensorflow/docs_src/performance/xla/operation_semantics.md
index 32f249c..217ab59 100644 (file)
--- a/tensorflow/docs_src/performance/xla/operation_semantics.md
+++ b/tensorflow/docs_src/performance/xla/operation_semantics.md
@@ -788,9 +788,7 @@ DynamicSlice extracts a sub-array from the input array at dynamic
 dimension: [start, start + size). The shape of `start_indices` must be rank ==
 1, with dimension size equal to the rank of `operand`.
 Note: handling of out-of-bounds slice indices (generated by incorrect runtime
-calculation of 'start_indices') is currently implementation-defined. Currently,
-slice indices are computed modulo input dimension sizes to prevent out-of-bound
-array accesses, but this behavior may change in future implementations.
+calculation of 'start_indices') is currently implementation-defined.
 
 <b> `DynamicSlice(operand, start_indices, size_indices)` </b>
 
@@ -847,9 +845,7 @@ is updated.
 The shape of `start_indices` must be rank == 1, with dimension size equal to
 the rank of `operand`.
 Note: handling of out-of-bounds slice indices (generated by incorrect runtime
-calculation of 'start_indices') is currently implementation-defined. Currently,
-slice indices are computed modulo update dimension sizes to prevent out-of-bound
-array accesses, but this behavior may change in future implementations.
+calculation of 'start_indices') is currently implementation-defined.
 
 <b> `DynamicUpdateSlice(operand, update, start_indices)` </b>